Should we let AI read our break-up emails?
Summary
Tom and new co-host Jessica Maddox discuss the hype around AI apocalyptic predictions, noting that these "p-doom" estimates are based on vibes rather than science. They also examine Meta’s new AI agent, Muse, which Tom tested by granting it access to his email, highlighting the tool’s convenience alongside serious privacy and security risks. Finally, the pair explore the rise of microdramas, or "verts," analyzing how this short-form content is reshaping Hollywood and reflecting our shrinking attention spans.
Topics discussed
Sponsors and show introduction
Host changes and welcome for Jessica Maddox
Jessica Maddox's background in meme studies
AI apocalypse fears and the 10% probability claim
Defining P-Doom and varying expert estimates
Critique of AI risk predictions as 'vibes'
Tech CEOs lobbying for regulation before elections
Human agency in AI development and Meta's Muse launch
Tom's experience setting up Meta's Muse agent
Granting Muse access to personal email and data
Privacy concerns and invasive data requests
Using Muse to negotiate subscription cancellations
Impact of AI agents on ticket buying and human connection
Infrastructure challenges and the 'annoying' internet
AI agents and the erosion of personal gestures
Commercial incentives and biased recommendations
Meta's privacy promises and data training defaults
Security vulnerabilities and data privacy incidents
Sponsors and transition to microdramas
Rise of vertical microdramas in Hollywood
Common tropes and titles in microdramas
Personal experience watching microdramas
Comparison to AI satire and industry growth
Viewership stats and vertical film festivals
AI slop replacing human actors in microdramas
Monetization models and paywalls for microdramas
Media theory and changing attention spans
Closing remarks and subscription call to action
Listen ad-free on Castria